Data Shows Internet Inaccessibility During Iranian Election

In the lead up to the presidential elections, Iranians grappled with a more restricted Internet. We know the government slowed Internet speeds (again) in the days before the country’s June 15 elections because it acknowledged as much. But research by ASL 19 and Herdict indicates the problems this month were far more extensive than simply slow connections.

Herdict and ASL 19, an anti-censorship advocacy organization, recently partnered to monitor Iranian censorship. Between June 1 and July 1, Herdict received 3,533 reports from Iran. Additionally, ASL 19 surveyed 555 Iranian Internet users about site accessibility in the country beginning on June 11. The survey data shows Iranians experienced extremely poor access to the Internet and various web services before the election, with a spike in inaccessibility on election day. ASL 19 posted summaries of each day’s findings on its blog.

Comparing Herdict data with ASL 19′s data allows us to see a clearer picture of the scope of the disruption to and restrictions on Internet freedom in Iran. Mohammad Hassan Nami, Iran’s minister of communications and information technology, told Tasnim News Agency that the restrictions were part of “security measures taken to preserve calm in the country during the election period,” according to Radio Free Europe. However, our data shows a far broader set of restrictions on Internet usage.

 .ir versus non-.ir Websites

Unsurprisingly, Iranians reported greater difficulty with foreign sites than local ones.  Between June 1 and July 1, Herdict received 2,283 accessible reports and 1,250 inaccessible reports from Iran, which is 35 percent inaccessible. Only 73 of the inaccessible reports pertained to websites using the Iranian top-level domain .ir; the remaining 1,177 inaccessible reports came from non-.ir websites (e.g., facebook.com, youtube.com, etc.). The percentage of non-.ir websites reported as being inaccessible decreased just prior to the election and has remained lower than pre-election levels. Between June 1 and June 10, 47 percent of non-.ir website reports were inaccessible. Between June 11 and June 15, the rate decreased to 39 percent. After the election, between June 16 and July 1, 40 percent of reports on non-.ir sites were inaccessible.

Herdict Inaccessibility Reports for non-.ir Sites

ASL 19 Survey Data on Accessibility of .ir and non-.ir Websites Between June 11-July 1

ASL 19 survey data also shows users experiencing more difficulty with non-.ir websites compared to .ir websites. Between June 11 and July 1, 22 percent of users reported normal access to .ir websites, compared to six percent who reported normal access to non-.ir websites. Herdict data from the same time period shows 90 percent of .ir site reports labeled accessible, compared to 60 percent of non-.ir site reports. Both Herdict and ASL 19 data show that far more users reported at least some problem accessing non-.ir sites compared to .ir sites.

It is useful to look at some specific, popular sites that both Herdict and ASL 19 tracked during the pre- and post-election period. Below we provide some details about Facebook, Twitter, and YouTube’s accessibility in Iran.

Facebook

249 Herdict reports from Iran were about Facebook. Of those, 49 percent reported the site inaccessible and 51 percent accessible. ASL 19′s survey data showed 18 percent had no access to the site, 70 percent experienced some difficulties accessing the site, and 7 percent had normal access.

Twitter

41 Herdict reports were for Twitter, and 88 percent recorded inaccessibility. This is comparable with ASL 19 survey data, in which only 3 percent of users reported normal access to the site. Nearly all other respondents who used Twitter reported either difficulty accessing the site or no access at all.

YouTube

69 Herdict reports were for YouTube. Of those, 61 percent reported the site inaccessible and 39 percent accessible.

Herdict Accessibility Reports Between June 1 and July 1

ASL 19 Survey Data on Website Accessibility Between June 11 and July 1

Ultimately, both datasets appear to say a very similar thing: very few people in Iran had normal, unfettered access to these popular sites.  Where there are differences, it is likely a result of the different scales that Herdict and ASL 19 used.  Herdict uses a binary accessible or inaccessible measure, whereas ASL 19 asked people to rate the quality of their access ranging from no access to normal access.  For these sites, that difference this can lead to somewhat divergent numbers.  If someone in Iran tries to access YouTube and service is so slow that it makes watching videos impossible, they might report that as inaccessible on Herdict but limited access for ASL 19′s survey.

Herdict also received inaccessibility reports for the BBC and BBC Persian, CNN, Reddit, Meyar News, Google, Amnesty International, and Herdict partner Citizen Lab.

Although Iran relied heavily on Internet filtering heading into this election, there may be some good news on the horizon.  Shortly after his election, President-elect Hassan Rouhani acknowledged that Internet filtering doesn’t work and called social networking sites, “a welcome phenomenon.” Whether that means fewer inaccessibility reports will flow from Iran remains to be seen.